Resumé

This article addresses the question of legitimacy, both in a general theoretical perspective and more specifically in relation to policy making in the European Union. It is argued that Jürgen Habermas' recent theories of discursive will formation provide a useful framework for understanding the current problems of legitimacy in European policy making. The main proposition is the central connection between communication and legitimacy, and it is argued that improving the communicative character of policy making should be a central aspect of the democratization efforts in Western Europe.

Introduction

States today face tremendous challenges trying to deal with the consequences of transnational interdependence, especially in the economic field. Many states suffer from severe financial problems, and efficiency problems urgently need to be addressed. One challenge is how to undertake the necessary changes in a socially acceptable way. We can conceptualize this as the legitimacy question. In this article I consider the relevance of Jürgen Habermas' more recent views on discursive will formation for addressing this core question of contemporary policy making. In a context where traditional institutional solutions such as representative parliamentarianism seem insufficient, as in the context of supranational decision making, Habermas' approach provides a flexible and promising guide for future institutional reform.

I argue for the importance of discursive will formation in framing policy that is rational and at the same time has good prospects of meeting public acceptance. The article proposes that the connection between communication and legitimacy is key. Accepting this proposition has implications for the

---

* Kerstin Jacobsson, Department of Sociology, Uppsala University, Box 821, 751 08 Uppsala, Sweden.

organization of policy making and imposes certain demands on the
relationship between the political system and civil society.

The legitimacy problem arises in the borderland between state and society and must be resolved there. Both "pragmatic" approaches (wanting to substitute democratic legitimacy for instrumental outcomes, such as welfare) and "formalistic" approaches (where legality rather than legitimacy is stressed) tend to miss the central problem. The question of legitimacy is subtle. It involves trust which can neither be commanded nor "bought." I argue that improving the communicative character of policy making, should be a central aspect of the democratization efforts in Western Europe. This is also the case with the European Union (EU) and its particular system of governance. In this view, the nature of the EU's legitimacy problem is general and not specific to the EU. But the particular construction of the EU and its policy making process tends to make the problem more acute. Thus, while the theoretical approach suggested here is useful in more general legitimacy analyses, it may be of particular value in the analysis of the European integration project.

The first part of the article outlines a theoretical perspective on collective will formation in the constitutional state, drawing mainly on Habermas. The second part explores the relevance of this perspective in addressing the question of legitimacy in European politics. However, I want to emphasize that the aim of this article is not to conduct an empirical analysis along these lines but to suggest an approach to addressing the legitimacy question.

Discursive Will Formation

The term "will formation" emphasizes the character of process and continual praxis of citizens in framing a collective course of action.¹ The reason I call the will formation discursive (or deliberative) is that it is conducted by means of public argumentation and discussion. It is through public debate among free and equal citizens that proposals can be justified or legitimated. Such a conception of democracy favors a specific (namely discursive) principle of justification instead of a particular organizational form. Institutionalization of this principle varies throughout history. This conception further emphasizes that collective choices should be deliberative and not simply fit citizen preferences (Cohen 1989, 22; Miller 1993). This is in contrast to a conception of collective choice as the aggregation of individual, non-deliberative preferences.

A common will, then, is not given from the beginning and "discovered," but is formed in a process of deliberation and negotiation where new perspectives may be taken into account, new information added, preferences changed, and so on. In fact, one of the functions of deliberative will

formation - of having to present reasons that must be acceptable to others to survive the deliberative process - is precisely to reshape preferences. As Cohen says: "Assuming a commitment to deliberative justification, the discovery that I can offer no persuasive reasons on behalf of a proposal of mine may transform the preferences that motivate the proposal" (1989, 24; cf also Miller 1993). While I might take my own preferences as sufficient reason to advance a proposal, I need to make the proposal acceptable to others if it is to survive their critical reflection and examination.

Reflective participants will be more aware of the possible particularity of their position. Whether it results in changed preferences or not, deliberation in any case promotes reflectively held preferences. It also functions to clarify lines of conflict and to foster a common view of the problem under consideration and of what is at stake, and thus the scope of disagreement can be reduced (Knight & Johnson 1994). While it is also true that clarified conflict lines may decrease rather than increase the prospects for consensus, the likelihood of stable and lasting outcomes increases.

It should be stressed that collective will formation does not presuppose some basic consent but handles situations of both conflict and cooperation: it "refers to the stabilization of mutual behavioral expectations in the case of conflict or to the choice and effective realization of collective goals in the case of cooperation" (Habermas 1989, 145). It contains elements of both negotiation and argumentation which can lead to an outcome that all affected parties may agree is in their best interest. The outcome may well be - and often is - a compromise and/or a majority decision. However, voting is more rational and less arbitrary if preceded by deliberation. It then escapes mere decisionism and aggregation of pre-deliberative preferences.

The Role of Law

While communicative action and deliberative democracy are recurrent themes throughout Habermas' scholarly work, his thinking has recently developed towards a political theory which gives more weight to the complexity of contemporary society without giving up democratic ideals. The need to address the institutional questions of democratic will formation has led him to increasingly stress the role of law. In his most recent writings, Habermas locates the radical democratic principles inside rather than outside the political and legal institutions of the Rechtsstaat. He sees the law making procedure in the modern constitutional state as the institutionalization of collective, discursive will formation. However, the question of legitimacy and validity must be distinguished from the historically contingent question of how to institutionalize and organize will formation (Tuori 1989, 137; Habermas 1989, 154; 1994a, 13). Thus, agreeing on the need for

legitimization of law and on the role of the discursive principle of
justification in this respect does not lead to any immediate conclusions about
how to organize the political process in practice.

However, without an institutional framework that both makes deliberation possible and allows the formed opinion to be effectively channeled into the formal decision making system, discursive will formation is not viable.² Habermas himself has put more emphasis on the normative model of democracy and its general conditions and requirements than on the actual institutionalization of policy formulation, legislation, administration, etc. Departing from Habermas' general insights in this respect, the next step would be a discussion of concrete institutions and rules for will formation, but the discussion here concerns the usefulness of the general framework.

A constitutional state is characterized by its rule of law. Law is the medium by which peoples govern themselves, exerting self-determination by self-legislation. Law making procedures can be seen as the institutionalization of collective, discursive will formation by which the popular sovereignty is exercised by citizens prescribing laws for themselves. In the Rechtsstaat, Habermas stresses, "popular sovereignty is no longer embodied in an identifiable assembly of autonomous citizens. It draws back into the so to speak subject-less communication circuits of forums and associations" (Outhwaite 1994, 143).³ Here we see the importance Habermas ascribes to the public sphere or Öffentlichkeit (an analytical category that can comprise a plurality of autonomous public spheres). When citizens meet to discuss collectively relevant questions or to act together, they develop a communicative power that can and should exert influence on the political system.

Law is precisely the medium by which this communicative power is transformed into administrative power. By means of law, power gets binding force and can be effectively implemented. Law must be both compulsory and compelling in order to be effective and thus combine the threat of sanctions with an appeal to shared convictions (Outhwaite 1994, 140; also Bohman 1994, 910; Habermas 1994b, 136). Laws should be obeyed, not only because they are made according to correct law making procedures, but also because citizens feel that they deserve to be respected. I will return to this need for normatively grounded legitimization below.

Will Formation in Complex Societies

Modern societies are characterized by differentiation and complexity, and this makes it even more difficult to achieve the ideal of deliberation and agreement among free and equal citizens. Societal complexity fundamentally changes the conditions for popular sovereignty. Under societal complexity, where should deliberation take place, and in what institutional settings? And

to what extent can such deliberation govern society as a whole? If it cannot,
what would be the function of deliberation?

According to Habermas, we can say that law complements morality with regard to the problems of social coordination in large, complex societies with large-scale institutions (Habermas 1994b, 140).⁴ In such societies people cannot interact face to face. Societal coordination cannot be subjectdependent. Large-scale institutions such as the market economy can hardly be governed merely by moral norms. Law then functions as a means for organizing subsystems (such as the economy) that have become autonomous from the moral context of action (Habermas 1992c, 365). Law is one of the integrative mechanisms that remain in pluralistic and heterogeneous modern societies where tradition and religion have lost much of their integrative force (Bertilsson 1995, 12 ff).

Habermas distinguishes between different forms of societal integration: social integration by means of shared normsand system integration by means of functional coordination by market and administration and their steering media money and power. Law is also involved in regulating the economic and political subsystems. All three forms of societal integration can then be said to have a discursive character to the extent that they undergo a discursive process of legislation. Another question is where the initiative comes from, and whether law has discursive origins in the sense that it originates from deliberation among free and equal citizens about common concerns and not from indisputable dictates from the market, bureaucracy or technocracy. Consequently, the challenge is how to strike an effective balance between communicatively structured and functionally integrated spheres of social life, or in Habermas' terms, between system and life world (1992c). We could also frame the dilemma in Dryzek's terms as a battle between instrumental social control and discursive social and political interaction (1994, 88).

Contrary to authors such as Luhmann, Habermas does not think that the complexity and differentiation of contemporary society makes the question of normative legitimization outdated. The underlying problem is "the inadequate institutionalization of the principles of the Rechtsstaat rather than an overload imposed by these principles on a state activity which has become more complex" (Habermas quoted in Outhwaite 1994, 148). The question of efficiency should not be confused with the question of legitimacy ⁵ According to Habermas, the scarcest resource in complex societies is not the steering capacity of states or the productivity of the market but social solidarity. Habermas maintains that "the forces of social solidarity can today be regenerated in the forms of communicative practices of selfdetermination" (in Outhwaite 1994, 149).

Dryzek goes as far as to claim that increasing complexity in the modern
world actually facilitates the rise of discursive designs and encourages

communicative rationality. This is because it is very difficult for instrumental rationality and its corresponding steering strategies to cope with complex and controversial problems (1994, 53, 57 ff). The interrelationships that go with societal complexity may also help convert zero sum problems to positive sum relations (Dryzek 1994, 75).

According to Habermas, law is needed to mediate between the normative demands of communicative reason and the empirical facts of power and complexity. Law harbors a tension between validity or normativity and facticity which entails both an element of force and a need for recognition. Complex societies are ever more dependent on law, and this increases the need for legitimization.

Legitimacy vs. Legality

Since the constitutional state rules by law, a question arises as to how the legal forms of state authority can be legitimate. Positive laws can be changed at any time by the law maker, so what grants such changeable laws legitimacy? As opposed to Weber and legal positivists, Habermas does not see legality as an answer to this question, and he rejects the view that law is legitimate if enacted according to established legal procedures. Legitimization by procedure "gives no answer to the question of why the legitimating procedure itself, why the regular operation of the constitutionally sanctioned organs, why indeed the constitutional order itself is legitimate" (Habermas 1985, 102). The question is how the political order as such can be legitimate. The constitutional state must secure its legitimacy extra-\ega\\y.

Whether laws can be coercively enforced and whether political decisions can be respected even by those they disfavor depends on the legitimacy of the law making process itself. What lends the law making process acceptability is the extent of its democratic organization, that is the extent of its discursive and fair character. The fairness is guaranteed in the basic civil rights (such as freedom of opinion, expression, etc.), legal protection, and the rights of equal political participation. The discursive character in turn depends on the communicative exchanges with the public sphere by which discursively formed opinion can be channeled into the political system and transformed into actual policy. The legitimacy of law thus depends on the extent to which it preserves "an internal connection with the socially integrative force of communicative action" (Habermas quoted in Baynes 1995, 212).

Habermas stresses the link between human rights (which protect individual autonomy) and popular sovereignty (which protects collective autonomy): "The internal connection between popular sovereignty and human rights that we are looking for consists in the fact that human rights

state precisely the conditions under which the various forms of communication
necessary for politically autonomous law-making can be legally
institutionalized" (Habermas 1994a, 12 f).

Habermas maintains that "[o]nly those laws may claim legitimacy that meet with the agreement of all citizens in a discursive law-making process that is itself legally constituted" (quoted in Bohman 1994, 911 f). In the discursive process, the issues are subjected to the principle of universality of interests, that is to a test of whether the outcome may be in the interest of all affected. Only public deliberation can assessand justify the reasonableness of proposals and ideas. In other words, the source of legitimacy is the process of will formation as such, i.e. deliberation itself. In Manin's words: "a legitimate decision does not represent the will of all, but is one that results from deliberation of all" (1987, 352; cf Miller 1993, 77; Bohman 1994, 925).⁷ Here the emphasis is on deliberation, not unanimity.

The Public Sphere

The public sphere should not be conceived of as a specific institution but rather, in Habermas' recent terms, as "a network for the communication of contents and the expression of attitudes, i.e. of opinions, in which the flows of communication are filtered and synthesized in such a way that they condense into public opinions clustered according to themes" (quoted in Outhwaite 1994, 147). Public opinion is thus the result of communication, and not of the aggregation of individual opinions in surveys.

The role of the public sphere is not just to mediate between civil society and the state regarding concerns and needs, but to keep a critical eye on the state and its exertion of power. By means of communication and interaction, central issues are developed and thematized in the Öffentlichkeit, and can then be passed on to the formal political system and channeled into the political process. It is important that any type of issue can be taken up for public deliberation.

However, a separation between social power and political power, between the public sphere and the political system, needs to be maintained. Contrary to the state, the actors in the public sphere do not have access to means of enforcing their proposals. But then again, they are neither accountable for the proposals they forward, nor do they have to take into account all aspects of problems which formal decision makers cannot avoid (cf Eriksen 1994, 381). But even if only formally organized bodies can make the decisions in the end, "they must remain open and sensitive to those communicative flows of autonomous public spheres where formation of beliefs, issues and arguments is not yet integrated into the institutionalized decision-making processes . . . Institutional political will-formation would destroy the basis

for its own rational functioning if it were to clog the sources of spontaneity
inherent in an antecedent public sphere" (Habermas 1989, 154).

Thus, rational collective will formation puts certain requirements on the relations between civil society and the state, so that the free-floating themes and ideas developed in the former can effectively influence the formal decision making. It also requires that civil society functions in such a way that a rational public opinion can be formed. Here such aspects as the information available, the quality of education, and the scope of and precision with which the controversial issues are articulated play a role (cf Outhwaite 1994, 138). In practice we know that unequal power relations, unequal access to debate fora, varying skills in expressing oneself, etc. constrain discursive will formation. Thus, the need for democratization also applies to civil society itself. In reality there are many constraining factors on a well-functioning deliberative process: the time frame within which decisions have to be made, the size and complexities of societies, the number of people affected by decisions and thus legitimate participants in discourse, the asymmetries of competence and of availability of information, the varieties in motivations to participate, to name just a few.

However, the rationality of an outcome as well as its legitimacy depend on the conditions under which it was reached. It is therefore worth striving for optimal conditions for deliberation, although it is important to be realistic about the possibilities in practice. A discursive theory of political legitimization departs from an ideal of discourse and discursive conditions and evaluates actual situations according to this standard.⁸ As a minimum, adequate and publicly available information - to avoid distorted beliefs - and the guarantee of equal access to participation - to avoid agreements reached by forces other than the force of the best argument - are needed. Moreover, unrestricted communication and free expression of opinions are important so that good ideas and arguments are not hindered. Also, an appropriate institutional framework is needed to enable free, reasoned deliberation among equals in practice. Whether the dispute will finally be settled by a majority vote or by unanimity remains in each case an empirical question. However, the legitimacy of the outcome is in any case granted by the preceding deliberative procedure.

As mentioned above, Habermas now locates his theory of democracy within the existing institutions of the constitutional state. One may ask if Habermas has come to content himself with the increasingly formal democracy of today. While there is no doubt that political and legal institutions are crucial for societal integration in today's complex societies, one must ask how these institutions function in practice. Despite his emphasis on the need for flow of communication between the public sphere and the political system, Habermas devotes little attention to what can actually be done to secure this.

Politics have undergone tremendous changes over the last decades, and one may ask to what extent politics take place within the formal institutions of the state. Is Giddens right when he claims that the most important changes affecting people's lives today do not originate in the formal, political sphere and that the formal political sphere is only able to provide limited answers to these concerns? (Giddens 1994, 111). Is the role of law makers merely to adapt to changes that have already taken place outside of their control and thus to codify and attempt to legitimize dictates from elsewhere? Beck makes a useful distinction between politics (of the political system) and subpolitics (in the sense of autonomous subsystemic politics) (Beck 1994, 35; also Beck 1992). Subpolitics refer to the techno-economic development which can be said to fill a political function (in its society changing scope) but in a non-political (i.e. non-accountable) way. One obvious question is what are the implications of subpolitics - and of failing to take them into account - for the political and social order in Western societies in the long run?

However, to depart from the existing political system and its law making procedures and to improve its communicative character may be a fruitful - and realistic - path to both improving the quality of decisions and increasing their democratic legitimacy.

Institutional Prerequisites and Problems

When it comes to discursive will formation in practice, there are two interrelated aspects that need to be taken into account: institutional requirements for the formation of will and the substantive issues to be settled. External factors, such as economic interdependence, may pose constraints on institutions (their appropriateness and effectiveness) as well as on the concrete policy that is to be - and can be - worked out.

The point of a process of discursive will formation is to frame a general interest. During the process of deliberation, particular interests - i.e. interests that cannot be motivated in a way that appeals to all or at least most of the participants - are excluded or transformed into a general interest. In fact, it is only by means of discourse that particular interests can be identified and sorted out. Not even the proposals of interest groups are necessarily particular interests. It is thus problematic to take some interests as general and others as particular without sufficiently specifying the process and mechanisms by which these can be distinguished. Also, the borderline between political and technical issues, that is between issues that need to be settled discursively and those that can be handled by expertise alone, can only be drawn in a discursive process. No issues can be excluded from deliberation from the beginning, and all interests must be articulated and

assessed publicly in order to be legitimated. Moreover, to the extent that a trade-off between procedural fairness and efficiency needs to be made, this too requires fair discussion, agreement and review in order to be legitimate (cf Blaug 1996, 67).

In principle, I agree with Giddens when he states that "it is the aspect of being open to deliberation, rather than where it occurs, which is most important" (Giddens 1994, 115). However, deliberatively formed opinion cannot effectively influence formal decision making unless there is an institutional infrastructure to support it.⁹ This infrastructure may well be part of the institutions of the constitutional state. But this requires that these institutions function in such a way as to allow effective communicative exchange with civil society, and also that civil society itself functions in such a way that public opinion can be effectively formed. Accordingly, the institutional infrastructure must also include such things as the educational system and the mass media and, as mentioned above, the need for democratization also applies to civil society itself. It is very important that all affected citizens have access to arenas where issues are deliberated.¹⁰

Thus, while deliberative democracy in principle is fully in accord with representative democracy, its institutions might need to be reshaped and revitalized. Let us recall the basic principle of deliberative democracy: opinion to be formed before decisions, not afterwards, and thus legitimization through previous deliberation. The fact that formal democratic institutions exist does not necessarily imply that they function in a way that fosters a discursive will formation. What needs to be improved is the communicative character of the actual law making practices, i.e. their openness for inflow of discursively formed opinion.

Discursive will formation is just as important in the large-scale and complex policy making processes of today. Such a political order relies just as much on confidence and trust between electors and elected, policy makers and policy takers. At the national as well as the at European level and especially in their interrelations, an adequate flow of information must be ensured, publicity must be the norm, avenues of access to deliberation fora must be multiple and non-exclusive, etc. Above all, deliberative democracy (or in Giddens' words "dialogic democracy") is a matter of attitude - a willingness to listen and to take the perspective of others into account before opting for a special policy line. That attitude needs to be given institutional support to function at the level of political decision making. It also needs to be anchored in a more encompassing political culture.

Policy making in an internationalized environment (such as the EU framework) imposes institutional constraints on discursive will formation. The character of negotiations (between national interests) and its "culture" of non-publicity and anonymity of power and accountability limit the scope and space for public insight and debate. Presently there are few institutional

forms of supranational opinion, will and interest formation within the EU. The European Parliament is an obvious candidate, but so far it has neither legislative power nor the right to initiate proposals. Policy making in an internationalized context may also impose constraints on what issues can be discussed and proposed, that is on substantial policy. However, transnationalization may also entail new opportunities for discursive will formation. Subnational as well as supranational arenas for public debates may be created, and new technologies provide new opportunities for transnational communication.

Government and Communication

According to the perspective outlined above, parliamentary democracy is not only a system of governance (based on the steering media money and power), but also an arrangement for integrating different interests and opinions and for testing the legitimacy of those in power (cf Eriksen 1993, 51). States need both functional steering (regulation and coordination of action) and consensus-generating political dialogue.

The difficulties facing states today concern both their governing capacity and their communicative capacity. We may even speak of deficits in these respects. But, using Sweden as an example, this is not the same as to claim that Sweden as a nation-state suffers from a legitimacy deficit. My thesis would be that the probability of a future legitimacy deficit is as much a function of the state's communicative capacity as of its efficiency. Generally speaking, legitimacy is even more crucial for a political regime when it cannot guarantee efficiency. I have argued elsewhere that nation-states may have lost considerable efficiency and autonomy in the last few decades, but nonetheless it is to nation-states that citizens pledge their loyalty and allegiance (Jacobsson 1994). On the other hand, the European Union, to which Sweden and 14 other national states have surrendered part of their sovereignty, seems to suffer from a legitimacy deficit. If this is true, my argument would be that the legitimacy deficit is related to a lack of effective, communicative consent formation, and that an improvement in this area should be the main component of any strategy to overcome the problem.

Although I cannot develop the argument, I consider support for the EU to be instrumentally rather than normatively motivated, and accordingly the EU is more susceptible to efficiency problems than the national states. This may of course change in the future depending on developments within the member states as well as within the EU. However, attempts to substitute efficiency for legitimacy are problematic. The underlying assumption then is that favorable instrumental outcomes, such as welfare, will replace the need for democratic legitimacy. Or, according to the official Swedish position on

the EU, that democratic legitimacy can be provided nationally and the supranational elements can be legitimated instrumentally (cf Regeringskansliet 1995; Skr 1995/96: 30). But instrumentally motivated support for a political regime makes it extremely success-dependent and thus unpredictable and potentially unstable. Results are important for the support of a regime. They are necessary but not sufficient factors by themselves. Sometimes it is objected that politics are already too much talk and that it should concentrate on getting things done. The first question however is what things should get done, then how and why, and these matters can only be settled by dialogue. This is not to deny that deliberatively formed opinions also need to be effectively translated into actual policy and subsequently implemented.

States' lack of control and the communication deficit are closely connected. The less capable the state is of controlling the economy, the less open it may be for discursively formed opinion to influence its policy. There is less time and space for deliberation. When central policy decisions take the shape of informal consultations and agreements between key policy makers, such as party leaders, it to some extent replaces a broader deliberation and anchoring in Parliament and among the broader public.¹¹

An attempt to legitimize the fact that discussion does not take place before decisions are made is to claim that there is nothing to discuss. Thus, fewer and fewer matters are open to deliberation, and more and more they take the shape of "necessities" - not open to questions or discussion (cf Jacobsson 1995). Politicians are reduced to "bookkeepers of the great necessities" (Olsen drawing on Seip in Olsen 1995, 6). One illustrative example from Sweden is the October 1990 declaration of the intention to apply for EC membership. This was announced in a financial plan to stabilize the economy and had not been discussed before the announcement, neither at the Social Democratic party conference held only a few weeks before nor in the public sphere more generally. The process of mobilizing support for the membership came after the application was submitted.

Is it the case that politicians have difficulties making the hard decisions, and so to speak need the "great necessities" to be able to make decisions at all? If the political system is too late in undertaking changes, we could say that it suffers from a kind of rationality deficit. This phenomenon may have institutional roots, but it is far from obvious that the problem will be overcome by closing off the channels for public discursive will formation. A process of discursive will formation may show that the expectations and demands of the public are not exactly what political decision makers believe them to be. In any case a process of will formation functions to modify and revise initial and non-deliberative preferences.

Communication and Legitimacy

The model of discursive will formation defines the relation between political legitimacy and communication as follows: by means of public discourse the reasonableness and fairness of proposals can be assessed and thereby justified. Or put differently, legitimacy is produced through communicative action. Proposals are evaluated in relation to collective goals (also discursively set) and according to the general interest. Apart from the formal rights to equal participation, all this requires that public opinions are communicated and that public deliberation can somehow effectively influence the decision making.

In the Swedish law making system, bills are preceded by public inquiries and reports (by both parliamentarians and experts), circulation of bills to concerned bodies for comment, and public hearings, and this has allowed for considerable communicative exchange around the issues under consideration. This comparatively slow decision making system has proved very effective when it comes to implementation, since a high degree of consensus has been reached through previous dialogue phases. We can characterize Swedish democracy as a dialogue culture, something popular movements and their organizational culture and the principle of publicity and general access to official documents have also contributed to.

In Sweden, the needs of civil society have not only been channeled through political parties but also mediated by interest groups through corporate structures through which core groups have gained privileged access to the structures of decision making. It seems today as if political parties and sectorial interest groups are increasingly unable to catch and channel the movements and developments in society, such as new societal interests and conflicts. Instead, new lines of conflict are expressed through other channels than these. The Nordic referenda on EU membership, where lines of conflicts other than those between the political parties appeared, may serve as examples. In some cases, internal party divisions, for example between the party "establishment" and the "grass roots," are as big as the divisions between different parties (cf Brox 1995). The voters and the elected politicians seem to move in different directions.

Politicians as well as political analysts are bewildered. What is going on with the voters? Should we conclude that people are ill-informed about the state of things? Or are they simply spoiled? Would they change their minds if they got more information about the "necessities" of this or that policy? If so, it is still obvious that the communication flow between the public sphere and the political system is malfunctioning. Or are societal concerns and divisions today different than when the party structures and corporate structures were "cemented," and are these concerns not effectively communicated and channeled into the decision making structures? In my

view, common sense explanations, such as lack of public insight, reveal an
incapacity to analytically grasp the core of the current legitimacy problems.

I propose that (1) the communicative element of Swedish democracy (exemplified above) is connected with its traditionally high degree of legitimacy, and (2) to the extent that this communicative element is eroding (without being compensated for, at for instance the supranational level), legitimacy risks being eroded too. A more general proposition is that communicative policy making has the best prospects for legitimization, and that the democratization efforts in Western Europe today are best supported by a strengthening of this communicative element - both nationally and supranational^. Deficient communication between actors in civil society and the formal political system may result in a mismatch between societal interests and concerns and the policies that are put into place. This in turn may result in increasing divisions and an erosion of trust between policy makers and policy takers that may eventually express itself in the form of alienation from formal politics and other societal institutions and/or in various extra-parliamentary manifestations.

EU Policy Making - Non-Communicative or Selectively Communicative?

I will devote some extra attention to the problems of EU policy making in relation to discursive will formation for two reasons. First, the scope and importance of EU law making has been constantly increasing. Second, the EU seems to suffer from more severe legitimacy problems than the member states. My thesis - once again - is that the legitimacy deficit is closely related to a communication deficit.

It should be stressed that the communication deficit is not mainly a question of lack of information. Therefore, the conclusion of European policy makers that they must become better at informing the citizens about the need for certain policies is misguided. Information is one-way communication, while dialogue implies a reciprocal relationship. Adequate information is a prerequisite for serious deliberation, but here it is more important how open the issues are for deliberation. In other words, the central question is the openness to inflow from civil society and the mediation of interests and concerns of the life worlds of citizens in the process of formal policy making. Of course, this has to be done in combination with an adequate inflow of expert opinion.

A few things may need to be said about the specific features of the EU, even if this is not the right place to go into detail on this matter. The originality of the EU's institutional order can be attributed to its combination of interstatal and suprastatal features. The fact that national governments do

the law making in the Council gives the cooperative effort the character of intergovernmental cooperation. On the other hand, the Council does have suprastatal features because the laws passed by the Council take precedence over national laws and may have a direct effect on citizens in individual countries, national governments may be voted down through the practice of majority voting, and nationals may thus be bound by decisions opposed by their own representatives. The European Parliament, the Commission and the Court are supranational institutions. Consequently, there are two political arenas: one national and one supranational. The participation of citizens commonly takes place at the national level, even when political authority is exercised supranationally.

The project of a European Union is not met with the same support from the citizens as from the political elites in the member states (cf the referenda on the Maastricht Treaty and the participation rates in the European elections). One assumption is that citizens feel that the EU interferes in too many things. The member governments have responded with measures intended to unload some of EU's current tasks from its framework. Hence the emphasis on the principle of subsidiarity. Some of the new policy areas, like education and culture, that were introduced by the Maastricht Treaty are supposed to go back to national governments altogether. Another aspect of self-criticism is the incomprehensibility of EU decision making to the citizens, hence the stress on transparency.¹²

The notion of a democratic deficit is well-established in the political discourse and is mostly used to denote the fact that the European Parliament does not have legislative power, and that citizens can only exert indirect and partial control over the law making body (the Council) through the accountability of individual ministers to national parliaments. However, the problem with this definition is that it limits the democratic problem to a lack of legislative power in a supranational parliament, and thus that it forces democratization efforts to follow this line of logic. Currently we know that there is limited support for a fully federalized Europe, and consequently the risk is that efforts to resolve the current legitimacy problem will be hampered if strengthening the European Parliament is conceived as the only way to democratize the union.

I will therefore use the concept "democratic deficit" in a more general sense. One aspect is the lack of avenues for control and accountability at the European level despite the fact that legislation can be directly applicable and have a direct effect on individual legal entities in the member states. The other aspect of the democratic deficit - and the one that catches my attention here - is the difficulties with discursive will formation in European policy making. This is not just a matter of lack of supranational interest and will formation, but also the non-communicative relationship between national and EU decision making. Policy making takes on the diplomatic character of

negotiations between national interests, taking place behind closed doors and with a low level of publicity. The common knowledge and understanding that is developed among the participants in such a context of international negotiation, and from which the general public is excluded, further tend to limit the space for public debate and accountability. Even nationally the debate is constrained by such things as the government's interest in guarding its negotiation position and not annoying its cooperative partners by publicizing their positions.¹³ Moreover, from the citizens' point of view, the positions of the member states are ready-made preferences - named 'national interests' - that can or cannot be fulfilled. Citizens are excluded from the process of preference modification which is so central for discursive will formation and for acceptance of outcomes.

Another aspect of the democratic deficit is the non-discursive origin of EU legislation. The civil service (however politicized it may be) in the Commission has exclusive right to propose draft legislation. In this process, lobbyists are central actors passing on the needs and interests of some privileged societal actors, while other less organized and financially weaker societal groups cannot speak their case as forcefully. While it is true that an increasing amount and scope of legislation increases the need for legitimization, the core legitimacy problem is not that the EU puts its oar in too many rivers. Rather, it is the fact that its policy has not been worked out discursively. Often the draft legislation does not emanate from the concerns of civil society and thus it strikes citizens as alien dictates from the outside. Or, to use Habermas' distinction between system and life world, the policy worked out tends to be alien to the life worlds of citizens. One such example is the detailed regulation of cucumbers and strawberries, however rational it may be in the system perspective of market and administration. This standardization is publicly defended as a "consumer interest," but did European consumers propose it or was it dictated by the market logic of competition rules? According to my line of reasoning, the solution is not to invoke the subsidiarity principle when it suits the central authority, but rather to open up for initiatives from below. Concretely, this could be done by giving parliamentarians (national as well as European) the formal right to initiate proposals, instead of assigning this exclusive right to the Commission.

According to this line of logic, the present EU decision making process is not communicative enough. Another possible interpretation, however, is that the EU has managed well so far because of its de facto "porous" structures. Lobbying and expert committees - or what has been called organized interest self-representation - represent de facto avenues of influence (Andersen and Burns 1996). The challenge is to develop this porous relationship between civil society and the decision making system in open, non-exclusive ways that foster a general interest. We can recall the other

"point" of discursive will formation which was to frame rational policy. While proposals may make perfectly good sense to the interest group in question, they may be quite alien and in some cases even dysfunctional to the greater whole. For instance, while the detailed harmonization of technical standards make perfect sense for the smooth functioning of the Common Market, it has other implications that are considered important from other points of view.

A recent response to the lack of public support for the EU has been to emphasize the need for concrete results and visible advantages of the EU cooperation. This is part of the current governmental position in Sweden, where it is stressed that "[t]he EU and the results that are achieved in the cooperation must be democratically rooted and correspond to the expectations held by citizens of the EU" (Regeringskansliet 1995, my translation). The idea is that the cooperation will be accepted by citizens to the extent it results in valued outcomes such as employment and welfare. A "paternalistic" attitude, where citizens are "fed" results without participating in defining the overall goals, is a pragmatic but short-sighted understanding of the legitimacy problem.¹⁴ What citizens expect from the cooperation can only be defined by the citizens themselves and passed on through their intermediaries to the formal legislative process. Therefore, the priorities of the cooperation must be determined and its limits and borders must be drawn through a discursive process. Seeking correspondence to alleged citizen expectations is dubious for yet another reason: The claim that proposals correspond to the interests of a people, a nation, or a majority group contains an authoritarian element that we must be aware of. Touraine says that "we must reject the conceptions, or rather ideologies, that claim a government is democratic if it effectively responds to the population's needs" (1995, 1). This cannot be - and usually is not - the criteria for democratic government.

Western Europe's basic problem is not a lack of benevolence and good intentions by political leaders, but the fact that in the EU system of governance, citizens are by and large excluded from the discursive formation of political will. Or more precisely, the prospects for deliberation to really "make a difference" is limited and the relation between public debate and actual decision making is thus weak (cf Stenelo 1990, 349 ff). While the current stress on the need for increased publicity is well in line with the idea of discursive will formation, it needs to be combined with effective means of influence. Moreover, with a policy making process that permits exclusive access for some societal interests, there is no guarantee that the result will be advantageous for everybody or even for a majority. On the contrary, it is quite likely that the outcome will reflect the interests of those parties who had access to influence.

In summary, it may be true that the public lacks relevant information
about the EU project and that the cooperation currently suffers from

efficiency problems, but the root of EU's legitimacy problem is neither an information deficit nor a temporary lack of concrete results due to economic recession or ineffective decision making. Its root is in the noncommunicative nature of the legislation process, both nationally and supranational^.

How to Move On - Politically and Analytically?

Transnationalization processes have come to challenge the institution of the nation-state both as a locus of power and as a repository of sovereignty. How, then, should we respond to this challenge - analytically and politically? Instead of starting the analysis in existing organizational forms, it might be fruitful to start from the point of view of principles and regulative ideals. An institution may be justified from different points of view, and different ideals will shape institutions in different ways. Above, I have argued for the importance of the discursive principle of justification in framing policy that is both rational and legitimate. The questions are: What new institutional arrangements are needed to fulfill that principle? How to improve existing arrangements as regards their communicative character? The crucial thing here is a public sphere where communicative relations as well as open relations between civil society and the political system can develop to enable a flow of communication. How can a communicative practice be institutionally supported and politically and culturally encouraged?

When it comes to policy making at the European level, it may be even more fruitful to start with principles. The discussion of EU's organizational form - a federation, a confederation or something sui generis - has in many senses come to a dead end, analytically as well as politically. The same is true for the corresponding strategies for the democratization of the EU. A new approach could be to start with democratic principles and then see what organizational forms would be required to secure these principles. In Habermas' interpretation, popular sovereignty is based on three principles: a) the principle of legal protection of individual rights, b) the legality principle that guarantees political and legal control of administration, and c) a principled separation of state and society that secures the discursive justification of social power in the transformation into administrative power (Habermas discussed in Eriksen 1994, 390). In my interpretation, the EU needs to develop forms for the discursive justification of decisions, and these may not take the forms they did in the nation-state. The separation of social and political organization would need to be upheld while simultaneously securing an openness between them. As for the principle of democratic rights, one could think of a well-developed European citizenship (including

civic and political rights), laid down in a constitution and secured by a court. The alternative would be to improve the effectiveness of national citizenship to influence European policy making, all along the line of discursive will formation. The third principle would be democratic control and accountability, for which some mechanisms would need to be found. In neither case, however, would the institutionalization of popular sovereignty need to take the same forms as in nation-states.

Starting with the principle of discursive justification, priority would be given to finding procedural requirements for discursive will formation, as to secure free and fair access to agenda setting arenas, deliberation and decision making. Communicative relations between the formal decision making system and public spheres would need to be improved at all levels but especially at the supranational level. I propose this as a task for political theorists as well as activists in the coming years, especially in light of the ongoing revisions of the EU treaties. In my view, this is a realistic and viable path to improving the existing de facto system of governance, which lacks the usual features of democratic government but nonetheless has proved able to not only remain for decades but also to continually extend its sphere of influence.

Habermas' views on discursive will formation provide a flexible response to the need for democratization, precisely because this approach is not limited to particular organizational forms or strategies. Regardless of whether the EU cooperation develops in a federal, confederal or completely new direction, the principle of discursive justification may always guide institutional innovation and also improve current ways of functioning to secure their democratic legitimacy.

Summary

In this article, I have concentrated on general procedures of opinion formation and policy making. The procedural requirements of discursive will formation aim at ensuring that policy outcomes will be as acceptable to those affected as possible. Since all proposals must be explicitly supported by reasons, compete with alternative proposals under fair conditions, and since critical reflection on differing positions is encouraged, policy outcomes will be as legitimate and as rational as possible. Both the legitimacy of a policy outcome and its rationality are thus functions of the conditions under which it is produced.

I have asserted that in conjunction with the current steering problems of governments, increasing communicative problems have evolved. But overcoming governmentability problems and increasing efficiency in noncommunicative ways may prove counterproductive in the long run, for

instance in the perspective of social integration. We live in an age where nothing can be taken for granted - neither welfare, nor democracy. In such times, policy making needs increased citizen support, whatever novel forms participation may take. For the EU, which is a political order still under development, a top priority must be to include the public voice in a discursive process of policy formation.

---

NOTES

1. Of course it is reminiscent of Rousseau's volonté generate. This does not imply an accept of all Rousseau's views. On the contrary, our view of the crucial role of communication among citizens in forming a common will clearly differs from Rousseau's opinion that such communication is neither needed nor desired. Besides, to my mind the term "general will" is no more controversial than the term "general interest."

2. Often, the idea of deliberative democracy is associated with the institution of direct democracy and thus dismissed, since direct democracy is infeasible in large, complex societies. However, as Cohen stresses, direct democracy, which is not the only arrangement for deliberative democracy, may not even be a particularly good one. Cohen himself stresses the role of political parties in making deliberative democracy possible (1989, 30 ff). Cf also Habermas 1979, 186.

3. Habermas here differs from Rousseau who saw the general will as connected to an identifiable assembly of citizens. He also refutes Rousseau's claim that the general will can only be expressed through direct citizen participation and never through representation.

4. It is with the uncoupling of law and morality (with positive law) that the modern legitimacy problem arises. A discrepancy appears between what is legal and what is perceived as just and right (cf Bertilsson 1987).

5. Consent may be motivated by instrumental outcomes, habit, passivity and even coercion. Legitimacy, in contrast, has a normative element to it - then in the eyes of the citizens the political system deserves to be subjected to. Even if efficiency may be one basis for legitimacy, there is no automatic connection between the two. Without recognition there is no legitimate political - or for that matter social - order. Order can be upheld by force, but there are normative and social limitations to the use of force in modern societies. Power, thus, needs to be legitimated.

6. This claim is strong and Bohman himself restates is as follows: "A law is legitimate if it is agreed to in a participatory process that is fair and open to all citizens" (1994, 922). Cf also Cohen: "outcomes are democratically legitimate if and only if they could be the object of a free and reasoned agreement among equals" (1989, 22).

7. I find this view useful for practical purposes. While the goal of all deliberation is the broadest possible acceptance, unanimity may not be reached in practice. In Manin's opinion, Habermas emphasizes unanimity too much (Manin 1987, footnote 35). I think Habermas does so because he grounds his discourse theory in his theory of language, where he claims that the possibility of mutual understanding and accord is inherent in our capacity to communicate. I do not think it is necessary to accept that proposition in order to appreciate his view of discourse and democracy. Furthermore, also Habermas acknowledges that not all conflicts are solved by consensus. He explicitly consents to Manin's interpretation in Habermas 1992d, 446.

8. Habermas has introduced the notion of an "ideal speech situation" as a critical standard or a counterfactual ideal against which actual situations can be compared. He grounds this in his language theory which I will not go into here. Cf also Cohen's "ideal deliberative procedure" in Cohen 1989.

9. On the need for discursive institutions and designs, see Dryzek 1994. Contrary to Habermas' recent work and to the perspective outlined in this article, Dryzek does not see the discursive designs formalized in constitutions and formal rules. He sees law as an alternative institution to discursive democracy and as the opposite of communicative rationality (1994, 83), while I argue that discursive will formation can be institutionalized in the law making process.

---

9. On the need for discursive institutions and designs, see Dryzek 1994. Contrary to Habermas' recent work and to the perspective outlined in this article, Dryzek does not see the discursive designs formalized in constitutions and formal rules. He sees law as an alternative institution to discursive democracy and as the opposite of communicative rationality (1994, 83), while I argue that discursive will formation can be institutionalized in the law making process.

10. "All citizens affected" and "deliberation of all" apply to the principle that no one affected by a decision should be excluded from its deliberation. It does not imply that everybody must participate, which would of course be utterly unrealistic. It means that those who wish to participate must be able to do so. In practice, that often means representing others who do not wish to participate.

11. Examples from the Swedish context are economic "crisis agreements" in 1992 and the preparation of the application for EC membership in 1991.

12. On the post-Maastricht responses to the problem of legitimacy, see for instance Lodge 1994 a and b.

13. I have expanded on the constraints on deliberation nationally in relation to EU policy making in Jacobsson 1996.

14. Instead of paternalism, we may also frame it as "de facto guardianship" (cf Dahl 1994, 33).

REFERENCES

Andersen, S. S. & Burns, T. R. 1996. "The European Union and the Erosion of Parliamentary
Democracy: A Study of Post-parliamentary Governance," in Andersen, S. S. & Eliassen, K.
A. (eds), The European Union - How Democratic is it? London: Sage Publications.

Baynes, K. 1995. "Democracy and the Rechtsstaat: Habermas's Faktizität und Geltung," jn
White, S. K. (ed), The Cambridge Companion to Habermas. Cambridge: Cambridge
University Press.

Beck, U. 1992 [1986]. Risk Society. Towards a New Modernity. London: Sage.

Bertilsson, M. 1987. "Om legitimitet," in Bergryd, U. (ed) Den sociologiska fantasin: teorier
om samhället. Stockholm: Raben & Sjögren.

Bertilsson, M. 1995. "Inledning," in Bertilsson, M. (ed), Ratten i forvandling: Jurister mellan
stat och marknad. Stockholm: Nerenius & Santérus Forlag.

Blaug, R. 1996. "New Theories of Discursive Democracy: A User's Guide." Philosophy &
Social Criticism, 22(1), 49-80.

Bohman, J. 1994. "Complexity, Pluralism, and the Constitutional State: On Habermas's
Faktizität und Gellung," Law & Society Review, 28(4), 897-930.

Brox, O. 1995. "Etter 28. november: mot en ny konstitusjonsdebatt?" Samtiden, 1, 20-24.

Cohen, J. 1989. "Deliberation and Democratic Legitimacy," in Hamlin. A. & Pettit, P. (eds),
The Good Polity: Normative Analysis of the State. Oxford: Basil Blackwell.

Dahl, R. A. 1994. "A Democratic Dilemma: System Effectiveness versus Citizen
Participation," Political Science Quarterly, 109(1), 23-34.

Dryzek, J. S. 1994 [1990]. Discursive Democracy: Politics, Policy, and Political Science.
Cambridge: Cambridge University Press.

Eriksen, E. O. 1993. Grenser for staten? Om makt, penger og kommunikasjon som styringsinstrumenter.
Oslo: Universitetsforlaget.

Eriksen, E. O. 1994. "Politikk og offentlighet. Den demokratiske rettsstat mellom faktisitet og
gyldighet," Tidskrift for samfunnsforskning, 3, 370-395.

Giddens, A. 1994. Beyond Left and Right: The Future of Radical Politics. Cambridge: Polity
Press.

Habermas, J. 1979 [1976]. Communication and the Evolution of Society. London: Heinemann.

Habermas, J. 1987 [1969]. Towards a Rational Society. Cambridge: Polity Press.

Habermas, J. 1989. "Towards a Communication-Concept of Rational Collective Will
formation. A Thought-Experiment," Ratio Juris, 2(2), 144-154.

Habermas, J. 1992a. "Citizenship and National Identity: Some Reflections on the Future of
Europe," Praxis International, 12(1), 1-19.

Habermas, J. 1992b. Faktizität und Geltung. Frankfurt: Suhrkamp.

Habermas, J. 1992c [1981]. The Theory of Communicative Action. Volume 2. Cambridge:
Polity Press.

Habermas, J. 1992d. "Further Reflections on the Public Sphere," in Calhoun, C. (ed),
Habermas and the Public Sphere. Cambridge: The MIT Press.

Habermas, J. 1994a. "Human Rights and Popular Sovereignty: The Liberal and Republican
Versions," Ratio Juris, 7(1), 1-13.

Habermas, J. 1994b. "Postscript to Faktizität und Geltung," Philosophy and Social Criticism,
20(4), 135-150.

Ingram, D. 1993. "The Limits and Possibilities of Communicative Ethics for Democratic
Theory," Political Theory, 21(2), 294-321.

Jacobsson, K. 1994. "Post-nationell demokrati? Tankar om EU, nationalstaterna och
demokratin," Haften for Kritiska Studier, 2-3, 120-131.

Jacobsson, K. 1995. "Konstruktionen av 'det nödvändiga': EU-debatten som offentligt
samtal," Haften for Kritiska Studier, 2, 55-68.

Jacobsson, K. 1996. "Sweden and the EU Membership: The Tacit Transformation from
National State to Member-State," Paper, CORE Research Seminar on the Study of European
Integration, Krogerup.

Knight, J. & Johnson, J. 1994. "Aggregation and Deliberation: on the Possibility of Democratic
Legitimacy," Political Theory, 22(2), 277-296.

Lodge, J. 1994a. "Transparency and Democratic Legitimacy," Journal of Common Market
Studies, 32(3), 343-368.

Lodge, J. 1994b. "The European Parliament and the Authority-Democracy Crisis," Annals,
531,69-83.

Majone, G. 1994. "Decisions publiques et deliberation," Revue Francaise de Science
Politique, 44(4), 579-598.

Manin, B. 1987. "On Legitimacy and Political Deliberation," Political Theory, 15(3), 338-368.

Miller, D. 1993. "Deliberative Democracy and Social Choice," in Held, D. (ed), Prospects for
Democracy. Cambridge: Polity Press.

Olsen, J.P. 1995. "The Future of the Nation-State," Paper, Jubilee Symposium. Uppsala:
Department of Government.

Outhwaite, W. 1994. Habermas: A Critical Introduction. Cambridge: Polity Press.

Skr. 1995/96: 30 Regeringens skrivelse 1995/96: 30 om EU.s regeringskonferens 1996.

Touraine, A. 1995. "Democracy and History," Paper, The Second Vienna Dialogue on
Democracy, Vienna.

Tuori, K. 1989. "Discourse Ethics and the Legitimacy of Law," Ratio Juris, 2(2), 125-143.